Catching command-line errors using %x

Question

Whenever you want to execute something on the command line, you can use the following syntax:

%x(command to run)

However, I want to catch a

Answer 1

You need a mix of @Cam 's answer and @tonttu 's answer.

decent explanation of $? and others.

Edit: the domain http://blog.purifyapp.com is now in hands of a domain-squatter and scammer.

result = %x(command to run 2>&1)

unless $? == 0 #check if the child process exited cleanly.
    puts "got error #{result}"
end

Answer 2

You might want to redirect stderr to stdout:

result = %x(command to run 2>&1)

Or if you want to separate the error messages from the actual output, you can use popen3:

require 'open3'
stdin, stdout, stderr = Open3.popen3("find /proc")

Then you can read the actual output from stdout and error messages from stderr.

Answer 3

So this doesn't directly answer your question (won't capture the command's output). But instead of trying begin/rescue, you can just check the exit code ($?) of the command:

%x(command to run)
unless $? == 0
   "ack! error occurred"
end

Edit: Just remembered this new project. I think it does exactly what you want:

https://github.com/envato/safe_shell

Answer 4

Here's how to use Ruby's open3:

require 'open3'
include Open3

stdin, stdout, stderr = popen3('date')
stdin.close

puts
puts "Reading STDOUT"
print stdout.read
stdout.close

puts
puts "Reading STDERR"
print stderr.read
stderr.close
# >> 
# >> Reading STDOUT
# >> Sat Jan 22 20:03:13 MST 2011
# >> 
# >> Reading STDERR

popen3 returns IO streams for STDIN, STDOUT and STDERR, allowing you to do I/O to the opened app.

Many command-line apps require their STDIN to be closed before they'll process their input.

You have to read from the returned STDOUT and STDERR pipes. They don't automatically shove content into a mystical variable.

In general, I like using a block with popen3 because it handles cleaning up behind itself.

Look through the examples in the Open3 doc. There's lots of nice functionality.